Deep Noise Suppression Challenge datasets contain clean speech and noise clips for audio enhancement research. The repository includes inference scripts and an ONNX model for a baseline Noise Suppression Network. The dataset was uploaded by ltnghia on July 28, 2025.
Use Cases
- Training deep noise suppression models based on provided clean speech and noise clips.
- Benchmarking speech enhancement algorithms against the baseline Noise Suppression Network.
- Evaluating audio quality improvements using the challenge's standardized dataset.
Strengths
- Includes a baseline ONNX model and inference scripts for immediate evaluation.
- Dataset is associated with the Interspeech 2020 academic challenge, suggesting a research-oriented purpose.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- ltnghia via Hugging Face
- Collection Method
- Likely compiled for the Interspeech 2020 Deep Noise Suppression Challenge.
- Freshness
- Last updated 2025-07-28 14:34:35; freshness should be verified.